MADE BY GOOGLE, POWERED BY MACHINE LEARNING

The introduction of AI and ML has redefined the Google product experience, driving further investment in hardware development and software integration. As information flows become more complex, Google relies on ML and AI to raise the most relevant data at the most relevant times. Now, billions of users rely on Google Photos to manage and share their memories, utilize Google Maps to find the most convenient parking spot, and rely on Gmail to automatically draft responses to unanswered emails — entirely from a Google device.

A NEW PRODUCT STRATEGY

The evolution of machine learning (ML) has driven Google to reconsider how it optimizes product performance and delivers meaningful user experiences. “All of Google was built because we started understanding text and webpages,” CEO Sundar Pichai announced in his keynote address at Google I/O 2017, the company’s annual developer-focused conference. “The fact that computers can understand images and video has profound implications for our core mission.” [1] While working on consumer operations at Google, I witnessed the company shift its strategy from being mobile first to AI (artificial intelligence) first. Under this new strategy, Google has moved from turning phones into the ‘remote controls’ of ours lives to creating universally accessible products with smarter computing. [2] The strategy manifested most visibly in product development where further investment was made to build Google-branded hardware and leverage ML to make the technology as intelligent and intuitive as possible.

With a huge user base and access to information, Google is well positioned to create these ML-led technologies. Google’s mission is to organize the world’s information and make it universally accessible and useful. [3] Initially, this mission drove the company to create its well-known search engine and and an open source operating system (OS) for third-party hardware makers to power their products. While the approach is proven successful by the estimated 1.4 billion consumers using Android OS, Google is now building its own hardware products and integrating the hardware with its advanced AI and ML technology, notably marketed as the Google Assistant. [4] Pairing software and hardware development has enabled Google to scale faster and build a distinct brand among competitors. Within two years, the company launched the Pixel smartphone, the Google Home smart speaker, and the Daydream VR headset. Offering a suite of hardware products merged with ML has deepened the user’s integration with, and dependence on, the broader Google ecosystem. [5]

Source: Google Blog, The Keyword (latest product launch: The Google Home Hub)

The introduction of AI and ML has redefined the Google product experience and set the business up for continued success and expansion. Distinct to Google devices is the technology that gives users the power to easily and quickly access information. Whether from the home or in the palm of one’s hand, Google more effectively meets users where they are. As information flows become more complex, Google relies on ML and AI to raise the most relevant data at the most relevant times. Now, billions of users rely on Google Photos to manage and share their memories, utilize Google Maps to find the most convenient parking spot, and rely on Gmail to automatically draft responses to unanswered emails — entirely from a Google device.

HARDWARE BUILT FOR THE SHORT AND LONG TERM

The company’s investment in hardware has proven to be a worthwhile investment, sparking changes in the organization’s structure. Based on Google’s financial statements, ‘other revenues,’ which includes sales from hardware, showed a 49% increase year-over-year (YoY) during Q1 2017, amounting to nearly $4 billion in revenues. [6] To sustain this momentum in the short term, Google consolidated its hardware teams and chartered the group with the company’s AI first strategy. A more visible example of this approach is highlighted by the reintroduction of Nest, which offers internet-connected thermostats, smoke detectors, and security cameras, from Alphabet back into Google. As speculated prior to the announcement, pulling Nest closer has enabled tighter integration of Google’s AI and ML services with its hardware products and expanded its device offerings. [7]

As ML and AI continue to serve as competitive advantages for Google’s hardware, the company is also invested in remaining the industry leader of this technology for the long term. Google has dedicated an entire team, referred to as Google AI, to conduct research, find more opportunity to integrate ML and AI into products, and develop tools that broaden access to these technologies. [8] For example, the team built TensorFlow, an open-source ML library for research and production, to developers that want to quickly deploy computation across their platforms.

SET UP FOR SUCCESS

While there is no doubt AI and ML offers Google the opportunity to differentiate its products, the company will constantly need to prove the value of its technology. And as the the product line expands, the company will have to find creative ways to stand out among a crowded hardware market and will have to rely on the assumption that users are willing and able to integrate these products into their homes and lives. The large be on AI and ML raises many questions about the future of Google and the industry. Was it appropriate to invest heavily in hardware in order to grow and drive its AI first strategy? Or were there alternatives that could have placed Google on top of its competitors rather than among them in a crowded hardware market? Finally, in thinking about the long term impact of this technology, does Google have a role in humanizing AI and ML?

(word count 795)

 

[1]  Google Developers, “Google I/O Keynote (Google I/O ’17),” YouTube, published May 17, 2017, https://www.youtube.com/watch?v=Y2VF8tmLFHw, accessed November 2018.

[2]  Google blog. Google, “A personal Google, just for you,” https://googleblog.blogspot.com/2016/10/a-personal-google-just-for-you.html, accessed November 2018.

[3] Google homepage. Google, “About Google,” https://www.google.com/about/, accessed November 2018.

[4] Why Google’s wandering into hardware. (2017, May 2). Business Insider Intelligence. Retrieved from https://intelligence.businessinsider.com/post/why-googles-wandering-into-hardware-app-stores-reach-record-q1-tencent-plans-to-build-us-ai-lab-2017-5

[5] Google Home Hub defines its take on the smart home. (2018, Oct 11). Business Insider Intelligence. Retrieved from https://intelligence.businessinsider.com/post/google-home-hub-defines-its-take-on-the-smart-home-foghorn-update-adds-edge-machine-learning-for-iiot-anki-integrating-alexa-into-consumer-robot-2018-10

[6] Why Google’s wandering into hardware. (2017, May 2). Business Insider Intelligence. Retrieved from https://intelligence.businessinsider.com/post/why-googles-wandering-into-hardware-app-stores-reach-record-q1-tencent-plans-to-build-us-ai-lab-2017-5

[7] Alphabet considers giving nest a new home within google. (2017, Nov 30). Dow Jones Institutional News Retrieved from http://search.proquest.com.ezp-prod1.hul.harvard.edu/docview/1970480544?accountid=11311

[8] Google AI. Google, “Tools for Everyone,” https://ai.google/tools/, accessed November 2018.

Previous:

Using machine learning to bend the healthcare cost curve

Next:

Additive Manufacturing at GE in the Midst of a Turnaround (797 words)

2 thoughts on “MADE BY GOOGLE, POWERED BY MACHINE LEARNING

  1. Great article Farrah! You raise excellent points with Google’s entry and investment in the hardware market. With regards to Google’s investment and shift to AI and ML, I find it only natural that given Google’s positioning with access to a vast amount of data on a global scale. Google is well positioned to provide insights much faster than probably any other company.

    The interesting question you raise is that of hardware, and with the rise of IoT and reduce costs of sensors, hardware is just another method of reaching the user. Google could have elected to collaborate with another large hardware company, but with the intense competition in other areas, it would just make Google’s vision for AI harder to achieve. Other companies in the tech space are facing political and social challenges and need to maintain control of how they manage users (really, the world’s) data. For these reasons, I believe Google is on the right track in investing in hardware to maintain a certain level of control over shaping this mega-trend’s future, while also maintaining close collaboration and investments in the wider technology community.

  2. Awesome article. As Farrah mentions, Google is well poised to remain a leader in the ML/AI space due to the wealth of data and captive audience they have across all various facets of our life. The final question raised regarding whether Google should play a role in “humanizing” AI and ML is a particularly thought-provoking one. Humanizing the technology can mean a lot of different things, but for me, the one that means the most, particularly for a company who aims to make the world’s information “universally accessible” is to democratizing the basic technology so folk’s even understand how ML/AI works. I know that Google is investing in this because I have tried poking around on its’ Google AI website (https://ai.google/education/) marketed to both those new to coding and seasoned ML practitioners.

Leave a comment